[gmx-developers] Semantics of acceleration groups.

[ramon at juguete.quim.ucm.es]

I am not sure if I understand it. Do you mean that
the velocities stored in state->v[] are effective
group velocities? Or are the velocities used
for the scaling, measured relative to the group
velocities, the effective group velocities?

I guess that the acceleration is applied to the group
velocity, and included in the velocity of each accelerated
particle. But it is discounted when discounted when 
evaluating a thermostat or barostat, so that only the
part of the velocity not contributed by the acceleration
is used. 

Am I right?

> >Looking at the code, it seems that the scaling effect of barostats
> >and thermostat is not applied to the velocity of the accelerating group.
> >This velocity is discounted before performing any other operation.
> > 
> >Why?
> So that you get effective group velocities.
> This can in principle be used to compute transport properties.